Convolutional Neural Networks¶
Project: Write an Algorithm for a Dog Identification App¶
In this notebook, some template code has already been provided for you, and you will need to implement additional functionality to successfully complete this project. You will not need to modify the included code beyond what is requested. Sections that begin with '(IMPLEMENTATION)' in the header indicate that the following block of code will require additional functionality which you must provide. Instructions will be provided for each section, and the specifics of the implementation are marked in the code block with a 'TODO' statement. Please be sure to read the instructions carefully!
Note: Once you have completed all of the code implementations, you need to finalize your work by exporting the Jupyter Notebook as an HTML document. Before exporting the notebook to html, all of the code cells need to have been run so that reviewers can see the final implementation and output. You can then export the notebook by using the menu above and navigating to File -> Download as -> HTML (.html). Include the finished document along with this notebook as your submission.
In addition to implementing code, there will be questions that you must answer which relate to the project and your implementation. Each section where you will answer a question is preceded by a 'Question X' header. Carefully read each question and provide thorough answers in the following text boxes that begin with 'Answer:'. Your project submission will be evaluated based on your answers to each of the questions and the implementation you provide.
Note: Code and Markdown cells can be executed using the Shift + Enter keyboard shortcut. Markdown cells can be edited by double-clicking the cell to enter edit mode.
The rubric contains optional "Stand Out Suggestions" for enhancing the project beyond the minimum requirements. If you decide to pursue the "Stand Out Suggestions", you should include the code in this Jupyter notebook.
Why We're Here¶
In this notebook, you will make the first steps towards developing an algorithm that could be used as part of a mobile or web app. At the end of this project, your code will accept any user-supplied image as input. If a dog is detected in the image, it will provide an estimate of the dog's breed. If a human is detected, it will provide an estimate of the dog breed that is most resembling. The image below displays potential sample output of your finished project (... but we expect that each student's algorithm will behave differently!).
In this real-world setting, you will need to piece together a series of models to perform different tasks; for instance, the algorithm that detects humans in an image will be different from the CNN that infers dog breed. There are many points of possible failure, and no perfect algorithm exists. Your imperfect solution will nonetheless create a fun user experience!
The Road Ahead¶
We break the notebook into separate steps. Feel free to use the links below to navigate the notebook.
- Step 0: Import Datasets
- Step 1: Detect Humans
- Step 2: Detect Dogs
- Step 3: Create a CNN to Classify Dog Breeds (from Scratch)
- Step 4: Create a CNN to Classify Dog Breeds (using Transfer Learning)
- Step 5: Write your Algorithm
- Step 6: Test Your Algorithm
Step 0: Import Datasets¶
Make sure that you've downloaded the required human and dog datasets:
Note: if you are using the Udacity workspace, you DO NOT need to re-download these - they can be found in the /data folder as noted in the cell below.
Download the dog dataset. Unzip the folder and place it in this project's home directory, at the location
/dog_images.Download the human dataset. Unzip the folder and place it in the home directory, at location
/lfw.
Note: If you are using a Windows machine, you are encouraged to use 7zip to extract the folder.
In the code cell below, we save the file paths for both the human (LFW) dataset and dog dataset in the numpy arrays human_files and dog_files.
import numpy as np
from glob import glob
# load filenames for human and dog images
human_files = np.array(glob("/data/lfw/*/*"))
dog_files = np.array(glob("/data/dog_images/*/*/*"))
# print number of images in each dataset
print('There are %d total human images.' % len(human_files))
print('There are %d total dog images.' % len(dog_files))
Step 1: Detect Humans¶
In this section, we use OpenCV's implementation of Haar feature-based cascade classifiers to detect human faces in images.
OpenCV provides many pre-trained face detectors, stored as XML files on github. We have downloaded one of these detectors and stored it in the haarcascades directory. In the next code cell, we demonstrate how to use this detector to find human faces in a sample image.
import cv2
import matplotlib.pyplot as plt
%matplotlib inline
# extract pre-trained face detector
face_cascade = cv2.CascadeClassifier('haarcascades/haarcascade_frontalface_alt.xml')
# load color (BGR) image
img = cv2.imread(human_files[0])
# convert BGR image to grayscale
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
# find faces in image
faces = face_cascade.detectMultiScale(gray)
# print number of faces detected in the image
print('Number of faces detected:', len(faces))
# get bounding box for each detected face
for (x,y,w,h) in faces:
# add bounding box to color image
cv2.rectangle(img,(x,y),(x+w,y+h),(255,0,0),2)
# convert BGR image to RGB for plotting
cv_rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
# display the image, along with bounding box
plt.imshow(cv_rgb)
plt.show()
Before using any of the face detectors, it is standard procedure to convert the images to grayscale. The detectMultiScale function executes the classifier stored in face_cascade and takes the grayscale image as a parameter.
In the above code, faces is a numpy array of detected faces, where each row corresponds to a detected face. Each detected face is a 1D array with four entries that specifies the bounding box of the detected face. The first two entries in the array (extracted in the above code as x and y) specify the horizontal and vertical positions of the top left corner of the bounding box. The last two entries in the array (extracted here as w and h) specify the width and height of the box.
Write a Human Face Detector¶
We can use this procedure to write a function that returns True if a human face is detected in an image and False otherwise. This function, aptly named face_detector, takes a string-valued file path to an image as input and appears in the code block below.
# returns "True" if face is detected in image stored at img_path
def face_detector(img_path):
img = cv2.imread(img_path)
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
faces = face_cascade.detectMultiScale(gray)
return len(faces) > 0
(IMPLEMENTATION) Assess the Human Face Detector¶
Question 1: Use the code cell below to test the performance of the face_detector function.
- What percentage of the first 100 images in
human_fileshave a detected human face? - What percentage of the first 100 images in
dog_fileshave a detected human face?
Ideally, we would like 100% of human images with a detected face and 0% of dog images with a detected face. You will see that our algorithm falls short of this goal, but still gives acceptable performance. We extract the file paths for the first 100 images from each of the datasets and store them in the numpy arrays human_files_short and dog_files_short.
Answer: (You can print out your results and/or write your percentages in this cell)
from tqdm import tqdm
human_files_short = human_files[:100]
dog_files_short = dog_files[:100]
#-#-# Do NOT modify the code above this line. #-#-#
## TODO: Test the performance of the face_detector algorithm
## on the images in human_files_short and dog_files_short.
def face(image):
img = cv2.imread(image)
# convert BGR image to grayscale
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
# find faces in image
faces = face_cascade.detectMultiScale(gray)
# print number of faces detected in the image
print('Number of faces detected:', len(faces))
# get bounding box for each detected face
for (x,y,w,h) in faces:
# add bounding box to color image
cv2.rectangle(img,(x,y),(x+w,y+h),(255,0,0),2)
# convert BGR image to RGB for plotting
cv_rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
# display the image, along with bounding box
plt.imshow(cv_rgb)
plt.show()
def performance_face_detector_opencv(human_files_short, dog_files_short):
face_human = [int(face_detector(human_img)) for human_img in human_files_short]
ratio_human = sum(face_human)/len(face_human)*100
#print("ratio_human",ratio_human)
print ('{}% of faces detected in the first 100 images in human_files by OpenCV'.format(ratio_human))
face_dog = 0
for dog_img in dog_files_short:
if face_detector(dog_img):
img = cv2.imread(dog_img)
face(dog_img)
face_dog += 1
ratio_dog = face_dog/len(dog_files_short)*100
#print("ratio_dog",ratio_dog)
print ('{}% of faces detected in the first 100 images in dog_files by OpenCV'.format(ratio_dog))
performance_face_detector_opencv(human_files_short, dog_files_short)
We suggest the face detector from OpenCV as a potential way to detect human images in your algorithm, but you are free to explore other approaches, especially approaches that make use of deep learning :). Please use the code cell below to design and test your own face detection algorithm. If you decide to pursue this optional task, report performance on human_files_short and dog_files_short.
### (Optional)
### TODO: Test performance of anotherface detection algorithm.
### Feel free to use as many code cells as needed.
Step 2: Detect Dogs¶
In this section, we use a pre-trained model to detect dogs in images.
Obtain Pre-trained VGG-16 Model¶
The code cell below downloads the VGG-16 model, along with weights that have been trained on ImageNet, a very large, very popular dataset used for image classification and other vision tasks. ImageNet contains over 10 million URLs, each linking to an image containing an object from one of 1000 categories.
import torch
import torchvision.models as models
# define VGG16 model
VGG16 = models.vgg16(pretrained=True)
# check if CUDA is available
use_cuda = torch.cuda.is_available()
# move model to GPU if CUDA is available
if use_cuda:
VGG16 = VGG16.cuda()
print(VGG16)
Given an image, this pre-trained VGG-16 model returns a prediction (derived from the 1000 possible categories in ImageNet) for the object that is contained in the image.
(IMPLEMENTATION) Making Predictions with a Pre-trained Model¶
In the next code cell, you will write a function that accepts a path to an image (such as 'dogImages/train/001.Affenpinscher/Affenpinscher_00001.jpg') as input and returns the index corresponding to the ImageNet class that is predicted by the pre-trained VGG-16 model. The output should always be an integer between 0 and 999, inclusive.
Before writing the function, make sure that you take the time to learn how to appropriately pre-process tensors for pre-trained models in the PyTorch documentation.
from PIL import Image
import torchvision.transforms as transforms
def VGG16_predict(img_path):
'''
Use pre-trained VGG-16 model to obtain index corresponding to
predicted ImageNet class for image at specified path
Args:
img_path: path to an image
Returns:
Index corresponding to VGG-16 model's prediction
'''
## TODO: Complete the function.
## Load and pre-process an image from the given img_path
## Return the *index* of the predicted class for that image
return None # predicted class index
def VGG16_predict(img_path):
'''
Use pre-trained VGG-16 model to obtain index corresponding to
predicted ImageNet class for image at specified path
Args:
img_path: path to an image
Returns:
Index corresponding to VGG-16 model's prediction
'''
## TODO: Complete the function.
## Load and pre-process an image from the given img_path
## Return the *index* of the predicted class for that image
#Open jpg
img = Image.open(img_path)
# convert img to tensor, to give it as an input for VGG16
toTensor = transforms.ToTensor()
# human face jpg file width 250
# dog jpg file size ar various, and then resize/crop to 250
transform_pipeline = transforms.Compose([transforms.RandomResizedCrop(250),
transforms.ToTensor()])
img_tensor = transform_pipeline(img)
img_tensor = img_tensor.unsqueeze(0)
# move tensor to cuda
if torch.cuda.is_available():
img_tensor = img_tensor.cuda()
prediction = VGG16(img_tensor)
# move tensor to cpu, for cpu processing
if torch.cuda.is_available():
prediction = prediction.cpu()
index = prediction.data.numpy().argmax()
return index # predicted class index
(IMPLEMENTATION) Write a Dog Detector¶
While looking at the dictionary, you will notice that the categories corresponding to dogs appear in an uninterrupted sequence and correspond to dictionary keys 151-268, inclusive, to include all categories from 'Chihuahua' to 'Mexican hairless'. Thus, in order to check to see if an image is predicted to contain a dog by the pre-trained VGG-16 model, we need only check if the pre-trained model predicts an index between 151 and 268 (inclusive).
Use these ideas to complete the dog_detector function below, which returns True if a dog is detected in an image (and False if not).
def dog_detector(img_path):
## TODO: Complete the function.
index = VGG16_predict(img_path)
return (151 <= index and index <= 268) # true/false
(IMPLEMENTATION) Assess the Dog Detector¶
Question 2: Use the code cell below to test the performance of your dog_detector function.
- What percentage of the images in
human_files_shorthave a detected dog? - What percentage of the images in
dog_files_shorthave a detected dog?
Answer:
Percentage of first 100 images where humans detected as a dog: 0.0%
Percentage of first 100 images where dogs detected as a dog: 82.0%
### TODO: Test the performance of the dog_detector function
### on the images in human_files_short and dog_files_short.
human = (np.average([dog_detector(img) for img in human_files_short]) *100)
dog = (np.average([dog_detector(img) for img in dog_files_short]) *100)
print("Percentage of first 100 images where humans detected as a dog: {}%".format(human))
print("Percentage of first 100 images where dogs detected as a dog: {}%".format(dog))
We suggest VGG-16 as a potential network to detect dog images in your algorithm, but you are free to explore other pre-trained networks (such as Inception-v3, ResNet-50, etc). Please use the code cell below to test other pre-trained PyTorch models. If you decide to pursue this optional task, report performance on human_files_short and dog_files_short.
### (Optional)
### TODO: Report the performance of another pre-trained network.
### Feel free to use as many code cells as needed.
Step 3: Create a CNN to Classify Dog Breeds (from Scratch)¶
Now that we have functions for detecting humans and dogs in images, we need a way to predict breed from images. In this step, you will create a CNN that classifies dog breeds. You must create your CNN from scratch (so, you can't use transfer learning yet!), and you must attain a test accuracy of at least 10%. In Step 4 of this notebook, you will have the opportunity to use transfer learning to create a CNN that attains greatly improved accuracy.
We mention that the task of assigning breed to dogs from images is considered exceptionally challenging. To see why, consider that even a human would have trouble distinguishing between a Brittany and a Welsh Springer Spaniel.
| Brittany | Welsh Springer Spaniel |
|---|---|
 |
 |
It is not difficult to find other dog breed pairs with minimal inter-class variation (for instance, Curly-Coated Retrievers and American Water Spaniels).
| Curly-Coated Retriever | American Water Spaniel |
|---|---|
 |
 |
Likewise, recall that labradors come in yellow, chocolate, and black. Your vision-based algorithm will have to conquer this high intra-class variation to determine how to classify all of these different shades as the same breed.
| Yellow Labrador | Chocolate Labrador | Black Labrador |
|---|---|---|
 |
 |
 |
We also mention that random chance presents an exceptionally low bar: setting aside the fact that the classes are slightly imabalanced, a random guess will provide a correct answer roughly 1 in 133 times, which corresponds to an accuracy of less than 1%.
Remember that the practice is far ahead of the theory in deep learning. Experiment with many different architectures, and trust your intuition. And, of course, have fun!
(IMPLEMENTATION) Specify Data Loaders for the Dog Dataset¶
Use the code cell below to write three separate data loaders for the training, validation, and test datasets of dog images (located at dog_images/train, dog_images/valid, and dog_images/test, respectively). You may find this documentation on custom datasets to be a useful resource. If you are interested in augmenting your training and/or validation data, check out the wide variety of transforms!
import os
from torchvision import datasets
### TODO: Write data loaders for training, validation, and test sets
## Specify appropriate transforms, and batch_sizes
from PIL import ImageFile
ImageFile.LOAD_TRUNCATED_IMAGES = True
resize = 224
crop = 224
directory = "/data/dog_images"
# define transforms
train_transforms = transforms.Compose([transforms.Resize(resize),
transforms.CenterCrop(crop),
transforms.RandomHorizontalFlip(),
transforms.RandomVerticalFlip(),
transforms.RandomRotation(20),
transforms.ToTensor(),
transforms.Normalize([0.485, 0.456, 0.406],
[0.229, 0.224, 0.225])])
test_transforms = transforms.Compose([transforms.Resize(resize),
transforms.CenterCrop(crop),
transforms.ToTensor(),
transforms.Normalize([0.485, 0.456, 0.406],
[0.229, 0.224, 0.225])])
train_data = datasets.ImageFolder( directory + '/train', transform=train_transforms )
test_data = datasets.ImageFolder( directory + '/test', transform=test_transforms )
valid_data = datasets.ImageFolder( directory + '/valid', transform=test_transforms )
print(' Number of train images are: ', len(train_data))
print(' Number of test images are: ', len(test_data))
print(' Number of valid images are: ', len(valid_data))
trainloader = torch.utils.data.DataLoader( train_data, batch_size=32, shuffle=True )
testloader = torch.utils.data.DataLoader( test_data, batch_size=16 )
validloader = torch.utils.data.DataLoader( valid_data, batch_size=16 )
loaders_scratch = {}
loaders_scratch['train'] = trainloader
loaders_scratch['valid'] = validloader
loaders_scratch['test'] = testloader
testloader
class_names = train_data.classes
number_classes = len(class_names)
param_op_size = len(class_names)
print("num of classes:", number_classes)
#print("")
#print("class names: \n", class_names)
inputs, classes = next( iter(loaders_scratch['train']) )
for image, label in zip(inputs, classes):
image = image.to("cpu").clone().detach()
image = image.numpy().squeeze()
image = image.transpose(1,2,0)
# normalize image
image = image * np.array((0.229, 0.224, 0.225)) + np.array((0.485, 0.456, 0.406))
image = image.clip(0, 1)
fig = plt.figure(figsize=(12,3))
plt.imshow(image)
plt.title(class_names[label])
Question 3: Describe your chosen procedure for preprocessing the data.
- How does your code resize the images (by cropping, stretching, etc)? What size did you pick for the input tensor, and why?
- Did you decide to augment the dataset? If so, how (through translations, flips, rotations, etc)? If not, why not?
Answer:
1) The training, test and validation datas were loaded
2) DataLoaders for each of these sets of datas were created.
3) Then all image were resized to 224 pixel. To avoid overfitting, they were center cropped, then randomly horizontal / vertical flip / rotations was done.
4) The imahes are (224, 224, 3), so the inputs are bigger than the labs (28, 28, 1) for Mnist and (32x32x3) for CIFAR.
5) Most of the pre-trained models require the input to be 224x224 pixel images so it needs to match the normalization used when the models were trained. Normalization is done separately with the means of [0.485, 0.456, 0.406] and the standard deviations of [0.229, 0.224, 0.225].
(IMPLEMENTATION) Model Architecture¶
Create a CNN to classify dog breed. Use the template in the code cell below.
import torch.nn as nn
import torch.nn.functional as F
# define the CNN architecture
class Net(nn.Module):
### TODO: choose an architecture, and complete the class
def __init__(self):
super(Net, self).__init__()
## Define layers of a CNN
# convolutional layer (sees 224x224x3 image tensor)
self.conv1 = nn.Conv2d(3, 16, 3, padding=1)
# convolutional layer (sees 112x112x16 tensor)
self.conv2 = nn.Conv2d(16, 32, 3, padding=1)
# convolutional layer (sees 56x56x32 tensor)
self.conv3 = nn.Conv2d(32, 64, 3, padding=1)
# max pooling layer
self.pool = nn.MaxPool2d(2, 2)
# linear layer (sees 28x28x64 -> 500)
self.fc1 = nn.Linear(28 * 28 * 64, 500)
# linear layer (500 -> 133)
self.fc2 = nn.Linear(500, param_op_size)
# dropout layer (p=0.25)
self.dropout = nn.Dropout(0.25)
# batch norm
self.batch_norm = nn.BatchNorm1d(num_features=500)
def forward(self, x):
## Define forward behavior
# add sequence of convolutional and max pooling layers
x = self.pool(F.relu(self.conv1(x)))
x = self.pool(F.relu(self.conv2(x)))
x = self.pool(F.relu(self.conv3(x)))
# flatten image input --> 28 * 28 * 64 = 50176
x = x.view(x.size(0), -1)
# add dropout layer
x = self.dropout(x)
# add 1st hidden layer, with relu activation function
x = F.relu(self.batch_norm( self.fc1(x)) )
# add dropout layer
x = self.dropout(x)
# add 2nd hidden layer, with relu activation function
x = self.fc2(x)
return x
#-#-# You so NOT have to modify the code below this line. #-#-#
# instantiate the CNN
model_scratch = Net()
# move tensors to GPU if CUDA is available
if use_cuda:
model_scratch.cuda()
Question 4: Outline the steps you took to get to your final CNN architecture and your reasoning at each step.
Answer:
Summary:
The first layer has input size of (224, 224, 3) and the output size of 133 classes that should be in the last layer.
Convolutional layers and maxpooling layers, and the linear layer were added and also dropout layers to avoid overfitting and to produce a 133-dim output.
MaxPooling2D was choosen bacause it is usually used to down-sample in these type of classification problems.
Lot of convolutional layers were includeed for the more complex patterns in color and shape to help the model to detect.
Steps :
1) The first layer is a convolutional layer that takes (224, 224, 3) input size.
2) New layer has 16 filters, each with a height and width of 3, and filter to have stride of 1 pixel
3) In the next step, the layer has the same width and height as the input layer, so padding is used
self.conv2 = nn.Conv2d(3, 32, 3, padding=1)
4) A pooling layer is added to down-sample the input's x-y dimensions, by a factor of 2:
self.pool = nn.MaxPool2d(2,2)
5) Then a fully connected linear layer is used at the end to produce a 133-dim output.
6) A Dropout layer is also added to avoid overfitting.
7) So the structure would be : ([16, 3, 224, 224])
([16, 16, 112, 112])
([16, 32, 56, 56])
([16, 64, 28, 28])
([16, 50176])
([16, 500])
([16, 133])
(IMPLEMENTATION) Specify Loss Function and Optimizer¶
Use the next code cell to specify a loss function and optimizer. Save the chosen loss function as criterion_scratch, and the optimizer as optimizer_scratch below.
import torch.optim as optim
#Param definition
learning_rate = 0.01
### TODO: select loss function
criterion_scratch = nn.CrossEntropyLoss()
### TODO: select optimizer
optimizer_scratch = optim.SGD(model_scratch.parameters(), lr = learning_rate)
(IMPLEMENTATION) Train and Validate the Model¶
Train and validate your model in the code cell below. Save the final model parameters at filepath 'model_scratch.pt'.
def train(n_epochs, loaders, model, optimizer, criterion, use_cuda, save_path):
"""returns trained model"""
print("start training for {} epochs ...".format(n_epochs))
# initialize tracker for minimum validation loss
valid_loss_min = np.Inf
# exist save-file, load save file
if os.path.exists(save_path):
print("load previous saved model ...")
model.load_state_dict(torch.load(save_path))
for epoch in range(1, n_epochs+1):
# initialize variables to monitor training and validation loss
train_loss = 0.0
valid_loss = 0.0
###################
# train the model #
###################
model.train()
for batch_idx, (data, target) in enumerate(loaders['train']):
# move to GPU
if use_cuda:
data, target = data.cuda(), target.cuda()
optimizer.zero_grad()
# forward pass: compute predicted outputs by passing inputs to the model
output = model(data)
# calculate the batch loss
loss = criterion(output, target)
# backward pass: compute gradient of the loss with respect to model parameters
loss.backward()
# perform a single optimization step (parameter update)
optimizer.step()
# update training loss
#train_loss = train_loss + ((1 / (batch_idx + 1)) * (loss.data - train_loss))
train_loss += loss.item()*data.size(0)
######################
# validate the model #
######################
model.eval() # ---- set model to evaluation mode
for batch_idx, (data, target) in enumerate(loaders['valid']):
# move to GPU
if use_cuda:
data, target = data.cuda(), target.cuda()
output = model(data)
# calculate the batch loss
loss = criterion(output, target)
# update average validation loss
valid_loss += loss.item() * data.size(0)
train_loss = train_loss / len(loaders['train'].dataset)
valid_loss = valid_loss / len(loaders['valid'].dataset)
print('Epoch: {} \tTraining Loss: {:.6f} \tValidation Loss: {:.6f}'.format( epoch, train_loss, valid_loss ),end="")
## TODO: save the model if validation loss has decreased
# save model if validation loss has decreased
if valid_loss <= valid_loss_min:
print(' Saving model ...')
torch.save(model.state_dict(), save_path)
valid_loss_min = valid_loss
else:
print('')
return model
# train the model
model_scratch = train(20, loaders_scratch, model_scratch, optimizer_scratch,
criterion_scratch, use_cuda, 'model_scratch.pt')
# load the model that got the best validation accuracy
model_scratch.load_state_dict(torch.load('model_scratch.pt'))
(IMPLEMENTATION) Test the Model¶
Try out your model on the test dataset of dog images. Use the code cell below to calculate and print the test loss and accuracy. Ensure that your test accuracy is greater than 10%.
def test(loaders, model, criterion, use_cuda):
# monitor test loss and accuracy
test_loss = 0.
correct = 0.
total = 0.
model.eval()
for batch_idx, (data, target) in enumerate(loaders['test']):
# move to GPU
if use_cuda:
data, target = data.cuda(), target.cuda()
# forward pass: compute predicted outputs by passing inputs to the model
output = model(data)
# calculate the loss
loss = criterion(output, target)
# update average test loss
test_loss = test_loss + ((1 / (batch_idx + 1)) * (loss.data - test_loss))
# convert output probabilities to predicted class
pred = output.data.max(1, keepdim=True)[1]
# compare predictions to true label
correct += np.sum(np.squeeze(pred.eq(target.data.view_as(pred))).cpu().numpy())
total += data.size(0)
print('Test Loss: {:.6f}\n'.format(test_loss))
print('\nTest Accuracy: %2d%% (%2d/%2d)' % (
100. * correct / total, correct, total))
# call test function
test(loaders_scratch, model_scratch, criterion_scratch, use_cuda)
Step 4: Create a CNN to Classify Dog Breeds (using Transfer Learning)¶
You will now use transfer learning to create a CNN that can identify dog breed from images. Your CNN must attain at least 60% accuracy on the test set.
(IMPLEMENTATION) Specify Data Loaders for the Dog Dataset¶
Use the code cell below to write three separate data loaders for the training, validation, and test datasets of dog images (located at dogImages/train, dogImages/valid, and dogImages/test, respectively).
If you like, you are welcome to use the same data loaders from the previous step, when you created a CNN from scratch.
## TODO: Specify data loaders
loaders_transfer = loaders_scratch
(IMPLEMENTATION) Model Architecture¶
Use transfer learning to create a CNN to classify dog breed. Use the code cell below, and save your initialized model as the variable model_transfer.
import torchvision.models as models
import torch.nn as nn
## TODO: Specify model architecture
model_transfer = models.vgg19(pretrained=True)
if use_cuda:
model_transfer = model_transfer.cuda()
print(model_transfer)
Question 5: Outline the steps you took to get to your final CNN architecture and your reasoning at each step. Describe why you think the architecture is suitable for the current problem.
Answer:
Its pretty efficient, a technique, to use transfer learning in order to solve the problems related to computer vision. That's because once we train them, these models work very well for feature detectors for images they were not trained on. Here I'm using transfer learning to train a network that can classify the dog images and for this task, I'll use a VGG-16 and VGG-19 model from torchvision model archiv, which was already trained previously.
The whole point of using VGG architectures was to reduce the number of parameters in the convolutional layers and improve training time. The multiple variants of VGGNet (VGG16, VGG19, etc.) differ only in the number of layers in the network. VGG16 has 138 million parameters. The important point is that all the conv kernels are of size 3x3 and maxpool kernels are of size 2x2 with a stride of 2. The idea behind having fixed size kernels is that all the variable size convolutional kernels used in Alexnet (11x11, 5x5, 3x3) can be replicated by making use of multiple 3x3 kernels as building blocks.
(IMPLEMENTATION) Specify Loss Function and Optimizer¶
Use the next code cell to specify a loss function and optimizer. Save the chosen loss function as criterion_transfer, and the optimizer as optimizer_transfer below.
import torch.optim as optim
#Param definition
learning_rate = 0.01
import torch.optim as optim
criterion_transfer = nn.CrossEntropyLoss()
# for VGG 19
optimizer_transfer = optim.SGD(filter(lambda p: p.requires_grad,model_transfer.parameters()), lr=learning_rate)
(IMPLEMENTATION) Train and Validate the Model¶
Train and validate your model in the code cell below. Save the final model parameters at filepath 'model_transfer.pt'.
# train the model again with 10 Epochs
model_transfer = train(10, loaders_transfer, model_transfer, optimizer_transfer,
criterion_transfer, use_cuda, 'model_transfer.pt')
# load the model that got the best validation accuracy (uncomment the line below)
model_transfer.load_state_dict(torch.load('model_transfer.pt'))
(IMPLEMENTATION) Test the Model¶
Try out your model on the test dataset of dog images. Use the code cell below to calculate and print the test loss and accuracy. Ensure that your test accuracy is greater than 60%.
test(loaders_transfer, model_transfer, criterion_transfer, use_cuda)
import os
from torchvision import datasets
from PIL import ImageFile
ImageFile.LOAD_TRUNCATED_IMAGES = True
### TODO: Write data loaders for training, validation, and test sets
## Specify appropriate transforms, and batch_sizes
#Paramaters Setting
param_transform_resize = 224
param_transform_crop = 224
param_data_directory = "/data/dog_images"
def process_image_to_tensor(image):
''' Scales, crops, and normalizes a PIL image for a PyTorch model,
returns an tensor array
As per Pytorch documentations: All pre-trained models expect input images normalized in the same way,
i.e. mini-batches of 3-channel RGB images
of shape (3 x H x W), where H and W are expected to be at least 224.
The images have to be loaded in to a range of [0, 1] and
then normalized using mean = [0.485, 0.456, 0.406] and std = [0.229, 0.224, 0.225].
You can use the following transform to normalize:
'''
# define transforms for the training data and testing data
prediction_transforms = transforms.Compose([transforms.Resize(param_transform_resize),
transforms.CenterCrop(param_transform_crop),
transforms.ToTensor(),
transforms.Normalize([0.485, 0.456, 0.406],
[0.229, 0.224, 0.225])])
img_pil = Image.open( image ).convert('RGB')
img_tensor = prediction_transforms( img_pil )[:3,:,:].unsqueeze(0)
return img_tensor
# helper function for un-normalizing an image - from STYLE TRANSFER exercise
# and converting it from a Tensor image to a NumPy image for display
def image_convert(tensor):
""" Display a tensor as an image. """
image = tensor.to("cpu").clone().detach()
image = image.numpy().squeeze()
image = image.transpose(1,2,0)
image = image * np.array((0.229, 0.224, 0.225)) + np.array((0.485, 0.456, 0.406))
image = image.clip(0, 1)
return image
# In[161]:
# show test image
param_test_dog_image = '/data/dog_images/train/001.Affenpinscher/Affenpinscher_00001.jpg'
dog_image = Image.open( param_test_dog_image )
plt.imshow(dog_image)
plt.show()
# ### (IMPLEMENTATION) Write a Dog Detector
#
# While looking at the [dictionary](https://gist.github.com/yrevar/942d3a0ac09ec9e5eb3a), you will notice that the categories corresponding to dogs appear in an uninterrupted sequence and correspond to dictionary keys 151-268, inclusive, to include all categories from `'Chihuahua'` to `'Mexican hairless'`. Thus, in order to check to see if an image is predicted to contain a dog by the pre-trained VGG-16 model, we need only check if the pre-trained model predicts an index between 151 and 268 (inclusive).
#
# Use these ideas to complete the `dog_detector` function below, which returns `True` if a dog is detected in an image (and `False` if not).
# In[82]:
### returns "True" if a dog is detected in the image stored at img_path
def dog_detector(img_path):
## TODO: Complete the function.
index = VGG16_predict(img_path)
return (151 <= index and index <= 268) # true/false
(IMPLEMENTATION) Predict Dog Breed with the Model¶
Write a function that takes an image path as input and returns the dog breed (Affenpinscher, Afghan hound, etc) that is predicted by your model.
### TODO: Write a function that takes a path to an image as input
### and returns the dog breed that is predicted by the model.
# list of class names by index, i.e. a name can be accessed like class_names[0]
class_names = [item[4:].replace("_", " ") for item in train_data.classes]
def predict_breed_transfer(img_path):
# load the image and return the predicted breed
image_tensor = process_image_to_tensor(img_path)
# move model inputs to cuda, if GPU available
if use_cuda:
image_tensor = image_tensor.cuda()
# get sample outputs
output = model_transfer(image_tensor)
# convert output probabilities to predicted class
_, preds_tensor = torch.max(output, 1)
pred = np.squeeze(preds_tensor.numpy()) if not use_cuda else np.squeeze(preds_tensor.cpu().numpy())
return class_names[pred]
def display_image(img_path, title="Title"):
image = Image.open(img_path)
plt.title(title)
plt.imshow(image)
plt.show()
import random
from PIL import Image, ImageFile
for image in random.sample(list(human_files_short), 4):
predicted_breed = predict_breed_transfer(image)
display_image(image, title="Predicted: {}".format(predicted_breed) )
Step 5: Write your Algorithm¶
Write an algorithm that accepts a file path to an image and first determines whether the image contains a human, dog, or neither. Then,
- if a dog is detected in the image, return the predicted breed.
- if a human is detected in the image, return the resembling dog breed.
- if neither is detected in the image, provide output that indicates an error.
You are welcome to write your own functions for detecting humans and dogs in images, but feel free to use the face_detector and human_detector functions developed above. You are required to use your CNN from Step 4 to predict dog breed.
Some sample output for our algorithm is provided below, but feel free to design your own user experience!
(IMPLEMENTATION) Write your Algorithm¶
### TODO: Write your algorithm.
### Feel free to use as many code cells as needed.
def run_app(img_path):
## handle cases for a human face, dog, and neither
# check if image has juman faces:
if (face_detector(img_path)):
print("Hello Human!")
predicted_breed = predict_breed_transfer(img_path)
display_image(img_path, title="Predicted: {}".format(predicted_breed) )
print("You look like a ...")
print(predicted_breed)
# check if image has dogs:
elif dog_detector(img_path):
print("Hello Dog!")
predicted_breed = predict_breed_transfer(img_path)
display_image(img_path, title="Predicted: {}".format(predicted_breed) )
print("Your breed is most likley ...")
print(predicted_breed)
# otherwise
else:
print("Oh, we're sorry! We couldn't detect any dog or human face in the image.")
display_image(img_path, title="...")
print("Try another!")
print("\n")
Step 6: Test Your Algorithm¶
In this section, you will take your new algorithm for a spin! What kind of dog does the algorithm think that you look like? If you have a dog, does it predict your dog's breed accurately? If you have a cat, does it mistakenly think that your cat is a dog?
(IMPLEMENTATION) Test Your Algorithm on Sample Images!¶
Test your algorithm at least six images on your computer. Feel free to use any images you like. Use at least two human and two dog images.
Question 6: Is the output better than you expected :) ? Or worse :( ? Provide at least three possible points of improvement for your algorithm.
Answer: (Three possible points for improvement)
- more number of training images for every class, would improve the accuracy of the model
- more epochs (more processsing ability)
- improved training images would also lead to better accuracy
## TODO: Execute your algorithm from Step 6 on
## at least 6 images on your computer.
## Feel free to use as many code cells as needed.
## suggested code, below
for file in np.hstack((human_files[:6], dog_files[:6])):
run_app(file)
Testing the model for google downloaded images¶
my_files = np.array(glob("new_images/*"))
for file in my_files:
run_app(file)
